Chapman’s Central Australian desert sandstone flora revisited

Museums are important repositories of palaeontological specimens that in some cases include historical collections that may be overlooked in contemporary studies. Fossil leaf impressions in silicified sandstone (=silcrete) from the deserts of northern South Australia, historically called the “desert sandstone”, have been known since at least 1872. A small collection of leaf impressions (c. 40) from the desert sandstone were identified by Tate in 1896 as Northern Hemisphere genera such as Alnus, Bombax, Quercus and Magnolia, together with Australian genera. However, these fossils were only first described by Chapman in 1937, who identified some leaves as Banksia, Cinnamomum and Eucalyptus. Chapman also recognised Flindersia, Hakea, Hedycarya, Lomatia, Nothofagus, Persoonia and Tristania and assigned them principally to fossil and extant species previously reported from Paleogene–Neogene fossil localities in New South Wales and Victoria. Most fossil taxa were based on 1–2, rarely > 4 fossils. However, the assigned botanical affinity of these leaf fossils was justified inadequately and consequently the taxonomic entities require reassessment using modern concepts, as was the case for Banksia praegrandis, which was revised previously and the species transferred to a fossil genus as the new combination Banksieaeformis praegrandis. The Tate fossil collection was largely forgotten but was recently relocated in the South Australian Museum. In this report, these silcrete fossils are re-examined and their systematics reassessed. No support for Cinnamomum, Flindersia, Hedycarya, Nothofagus, Persoonia, Tristania and other genera amongst these fossils was sustained, though eucalypt-like leaves are confirmed, likely representing Corymbia and Eucalyptus s. str.